Keir Starmer’s ‘AI Opportunities Action Plan‘ finally acknowledges AI’s importance to economic growth – that is, a few months after Labour cancelled £1.3bn in AI funding committed by the previous government. Even so, the plan lacks support for British AI startups, the very companies it should be championing if it wants the UK to lead the field in the future. While the plan trumpets impressive-sounding initiatives like supercomputers, growth zones, and a National Data Library, beneath the surface it’s yet another framework designed for tech giants rather than the nimble, UK-run startups.
Instead of the government asking itself how it might get out of the way of innovation, instead it responded to each of Matt Clifford’s proposed initiatives by creating additional bureaucratic layers—from licensing and compliance frameworks to oversight councils. For one of Clifford’s most crucial recommendations—”Reform the UK text and data mining regime so that it is at least as competitive as the EU”— the government merely committed to “launch a consultation” rather than taking decisive action. Meanwhile, the global AI race moves ahead at breakneck speed, leaving British startups in a bureaucratic fog.
Regulatory burdens hit small companies the hardest. While US tech giants maintain armies of lawyers to navigate complex compliance frameworks, UK startups face an impossible choice: burn precious capital and time on regulatory compliance or risk falling offside. The government could solve this by creating practical regulatory sandboxes, controlled environments where startups can innovate without drowning in paperwork. For its AI policy, the government should take inspiration from the Financial Conduct Authority (FCA) Sandbox implemented in 2016, which helped British fintech start-ups Monzo and Revolut scale rapidly.
Tellingly, in his UCL speech unveiling the policy, Keir Starmer referenced two US companies, Anthropic and Cohere, rather than highlighting British innovation. While the promised expansion of supercomputing capacity is of fundamental importance, there’s no guaranteed access for small and medium-sized enterprises. Once again, we risk creating a system where established US players with deep pockets dominate the resources meant to foster innovation.
Perhaps most concerning for the feasibility of the plan is the fundamental contradiction between the government’s AI ambitions and its net zero targets. Britain’s electricity costs are already eye-watering: $368 per MWh compared to $184 in the US and $78 in China. The plan calls for “at least a 20x increase” in computing power and 500MW data centres in AI Growth Zones, yet offers no practical solution for powering these energy-hungry facilities under strict net zero constraints.
The path to make the UK an AI superpower isn’t complicated. Rather than creating more councils and launching endless consultations, the government could take immediate action to fuel the next generation of UK AI startups by raising S/EIS investment caps for tax relief schemes, enabling UK startups to raise the kind of capital they need to compete with their US counterparts. It could also accommodate AI startups by streamlining qualification rules and extending SEIS/EIS eligibility periods. Another helpful decision would be to reverse the recent hike in National Insurance – a measure that, as far as UK startups are concerned, makes it prohibitively expensive to hire homegrown talent.
As a British entrepreneur running what is both a small business and a cutting-edge AI startup, I see this plan as a modest step forward. However, it follows multiple backwards steps this government has taken via its broader economic policy. Rather than creating councils, launching consultations, and implementing compliance frameworks, what would truly benefit the UK’s founders and workforce is reducing the basic cost for companies to hire local talent and removing regulatory complexity.
AI is destined to become the operating system powering almost all digital goods and services. If we want the foundations of this new and powerful technological substrate to reflect our values rather than those of Beijing or San Francisco, it’s crucial that British-built companies can compete. Britain won’t become an AI powerhouse by government committee – but it can if the government begins to lift the tax burden on startups and remove obvious roadblocks to innovation.
Max Sinclair is the founder of Ecomtent
Read more: There’s a looming skills crisis in cloud security
